Project Summary

This research project is in progress. PCORI will post the research findings on this page within 90 days after the results are final.

One of PCORI’s goals is to improve the methods that researchers use for patient-centered outcomes research. PCORI funds methods projects like this one to better understand and advance the use of research methods that improve the strength and quality of comparative effectiveness research.

What is this project about?

In randomized clinical trials, or RCTs, researchers assign patients to different treatments by chance to compare the benefits and harms. RCTs often take place in research centers, and researchers have a high level of control over how patients receive treatment. One type of RCT is a cluster RCT. In a cluster RCT, researchers assign sites to different treatments by chance. Patients then receive the treatment assigned to the site.

A pragmatic cluster RCT, or a CRT, is a type of cluster RCT that takes place where patients typically receive treatment, such as health clinics. CRTs can help capture the real-world effects of treatment but determining if a treatment works can be hard. Patients who go to different clinics may differ in ways that affect results. Standard methods used to analyze data from CRTs don’t do a good job of taking such differences into account, which can lead to inaccurate results.

In this study, the research team is developing and testing methods to improve the design and analysis of CRTs. The new methods account for differences between groups after treatment assignment. The team is applying the new methods to a CRT study of treatment for opioid use disorders, or OUDs.

How can this project help improve research methods?

The new methods may help researchers when considering ways to improve the accuracy of CRT results.

What is the research team doing?

This project has two parts. In the first part, the research team is creating ways to figure out how groups of patients differ after treatment assignment. Then, the team is developing two new methods for analyzing data from CRTs. One method uses only findings from the study in the analysis. The other method includes findings from previous studies in the analysis, which may help researchers obtain accurate results.

In the second part, the research team is using a computer program to create test data that mimic data from a study on medicines for OUD. The team is using the computer program and test data to see how well the new methods work compared with standard methods.

Research methods at a glance

Design Element Description
  • To develop and disseminate generalizable causal inference methods to address post-randomization selection bias in the design and analysis of CRTs
  • To evaluate causal effects of increasing medication-based treatment of patients with OUD on patient-level acute care utilization
Approach Defining causal estimands based on principal stratification, Bayesian analysis, and frequentist estimation using generalized estimating equations 

Project Information

Fan Li, PhD
Duke University
New Causal Inference Methods for Cluster Randomized Trials with Post-Randomization Selection-Bias

Key Dates

November 2019
January 2024

Study Registration Information


Award Type
State State The state where the project originates, or where the primary institution or organization is located. View Glossary
Last updated: November 30, 2022